Release
This postmortem is for the most part a lot of thoughts on puzzle design in games in general, but lets first quickly look at a list of other things we're learning from for the next game.
We’re very grateful for the “overwhelmingly positive” response to the game, sitting at 97% positive right now and having earned back its cost already. I did expect a response in the mid to high 80s, but not in the high 90s because I’m well aware of many of the game’s shortcomings. Though I have to say, my own view is probably overshadowed by the bumpy development process more than what the actual result is. In the end the bigger problems have been glossed over and are not that visible anymore to the player.
The issues of Six Inches Under have been pretty obvious to us for a longer time, so there are not too many things I learned after release. The team internal postmortem is the far more important one to us because it helps us improve our work processes. One of our main problems for example was that we did way too many things that took a lot of effort without providing a proportional amount of fun. The reason I could make Supraland by myself was that I only picked low-effort/high-fun things.
Origin Story
If you read the SIU devblogs you already know how this project was not really planned to exist in such a way and we’ve just been improvising along. It was supposed to be a DLC for players who already know the abilities, so there were no introductions to new abilities. We made the switch to standalone (because nobody buys DLCs) and then shoehorned in lots of tutorials.
The game kind of had no creative head for the most time because it was just this little experiment and it felt more like a committee thing (which I think is terrible thing for any piece of entertainment!).
So the first 1.5 years of its development I personally was busy making the Crash DLC and then I needed a serious break. Having made Supraball, Supraland and Supraland crash in a row was just too much. Making games was the least interesting thing to me at that point. So I got a new hobby and started playing pool billiards day in and day out. After half a year of not doing any game development the drive slowly came back to me.
In spring 2021 I took over Six Inches Under and filled the void of direction the project had.
I tried to wrap up the game, polish up all the rough edges, but there were fundamental problems we could not fix without redoing everything, so we just accepted them. For example:
Backtracking
To create motivational backtracking, you need to plan things out very well and have the world open and interconnected. I know how to do it and it will be fine in Supraworld; it just didn’t work out well in Six Inches Under for various reasons. One being that it’s underground and the other one being, that the areas are all sidearms to Cagetown and too far away from each other, so we could not interconnect them.
Supraland was very open, every region was interconnected. That's how it will be again in Supraworld. And it's just much more of a feel good vibe when the sun is shining!
Gating
A big question for anyone designing a metroidvania is, how many gates you let the player walk past, that they can’t overcome yet. That is in general a thing that was not accounted for a lot in the planning of SIU and you can feel the lack of it. In Supraland I was very careful about planning this. I think you can barely plan it out fully; it's way too complex. A lot of things will always just happen during development.
One of my core philosophies is, that I wanna let you run past many future gates, but I don’t want you to see them. I think it's a mediocre design choice if you see all the future problems and then start speculating about which ability you will get lateron to overcome it. You all know that feeling, I’m sure.
A prime example from the first Supraland, when you get the float buckle and suddenly you remember all the metal tools you saw everywhere. That’s pretty much always one of the highlights for everyone. But also knowledge gates like the purple flies in your hometown. You could have used them all along, but only lateron you know how.
The other question is, if you go on an adventure, like your typical Zelda Dungeon, away from your hometown, should you be able to 100% that dungeon now or only later when you got all the abilities?
From Six Inches Under a takeaway of mine is that we should let you 100% it with whatever you get in that region. We are dangling too many carrots in front of your face, but you have no chance yet to get them. It does not spark joy for me.
A thing I also want to overcome in Supraworld is the uncertainty if you can solve something. Whenever people are not sure if they can solve a thing, they will just leave and go somewhere else. That's sad. I think it's a beautiful situation when you know you can solve something, but you have no clue how, but you can take on the challenge.
Mood geometry without purpose
I think we did no good job with the interior of cage town, because you can do very few interesting things there. It is mostly architecture that is supporting the atmosphere and setting, but it is not interwoven with puzzles. All the puzzles that exist there barely deserve the name, because they are just busy work like climbing buildings to get some gold or to carry burning matches around. But nothing that makes you go “AHA!”.
As a contrast, look at Blue Ville in the first game. Every single building, every single brick placement, is part of the puzzles in there. Puzzles that sometimes stretch across multiple buildings. There are no useless houses that are just there for the mood. Everything serves multiple purposes. If you removed half the buildings from cagetown, you would lose atmosphere, but the gameplay would ultimately be the same.
I think we did a good job with the design of the different floors, the idea of 3 tiered class system in a hamster cage and how the mood changes. Did you notice the music changes from floor to floor? If you didn't notice consciously, I'm sure you felt it subconsciously.
Anyway, I want every single thing to have a purpose and not fill the world with distracting details, and that's how we'll do it in Supraworld.
Too many Shops
The first game established having many shops everywhere and all of them sold their own stuff. 7 shops more or less. SIU once again has multiple shops but less; only 4. I think it’s too confusing. What I wanna do in the future is that we have the same shop in multiple places in the world. So no matter where you go, it’s the same content everywhere.
Combat
Enemies as puzzle elements were definitely appreciated by players. That is something that I wanna do A LOT in Supraworld.
We pretty much only put enemies into arenas which separate them from the rest of the game. The change to keeping enemies away from puzzles is a good thing, obviously.
Limiting enemies to just arenas feels kinda meh to me though. When you leave your safe zones it should feel a little threatening to add a certain atmosphere to the game. We need to bring that back.
That enemies require different approaches to kill them is also good and I think everyone appreciated that. We'll certainly go more into that direction.
The combat feeling itself however I don’t like so much. The feeling of hitting someone, dodging someone etc. All those details don’t feel as good as they did in the first game. It is a lot of detail work to get this stuff right and nobody on the SIU team was really responsible for it and that’s why it ended up unpolished. I’ll kneel into these details in the next game to make sure it feels satisfying.
I also want to bring coin drops back from enemies. At the same time the coins should be very valuable. It should be a much better motivation to kill enemies instead of just doing it because the next door doesn’t open. Every enemy you attack should invoke the feeling that it gets you closer to the next major upgrade, because it will drop coins.
Another problem to take on is how we mix 1. reappearing enemies in the wilderness, 2. enemies all dropping coins and 3. having a tight, engaging coin balance. I think the answer is probably to make it way too grindy to wait for respawning enemies while the upgrade cost also goes up, so the roaming enemies don't provide enough.
Goals
The whole cave theme made it hard to give the player overarching goals.
We show you the exit sign, but you have no idea really how to get there. So we needed to guide you way more.
I love what we’re gonna do in Supraworld in that regard. The goals will always be visible in the distance, and it’s up to you to somehow manage to get there. In the cave you just could not make the goals visible in the distance, so leading the player through the main progression is a bit artificial and directionless. We often did this “find more abilities” quest. It sorta works, but it’s too ambiguous to be really engaging I think, because you have no real direction.
We’ve also gone through a couple of other goal types that didn’t work.
One I remember well was that to get on to the top floor you needed to look rich. So you needed to steal the golden (stomp) shoes from the bank vault to look appropriate. But it’s not the kind of idea players get by themselves; you need to actively tell them and that makes it boring.
My goal is always to make players feel clever for making decisions on their own and by figuring out the ways to execute on them. So the more specific and detailed the goals become, the more you kill that fun. AAA games basically have nothing of that left. They are so afraid of players getting stuck or lost, they just flat out tell you every step you need to make. It’s all absolutely unplayable for me at this point.
Supraworld will take the Supraland approach of having landmarks in the distance and those are your goal. All the little steps in between you have to come up with by yourself, based on the overarching goal. But it will be better than in Supraland, because you could not really visually see these goals most of the time, so it was still too ambiguous. In Supraworld the whole world is designed around that idea.
The following screenshot shows you a very early version of Supraworld. In the distance you see a wooden castle that you need to get on top of in order to..... nah, I'm not gonna spoil that much.
Zappy Gun
Anyway, instead of the golden shoes, the vault in Six Inches under now contains the Zappy gun. A tool that we know doesn’t add much to the game in a Supraland sense. In other metroidvanias you often get abilities that simply allow you to disable one type of barrier and nothing else. In Supraland though you expect all abilities to do all kinds of creative things that combine with other mechanics. The Zappy gun cannot deliver that. It mainly helps a lot in combat.
Abilities like the zappy gun just eliminate puzzle ideas. Imagine if in Supraland 1 you had a color gun that could color everything the way you needed it. This kills all the beautiful color puzzles. And that’s what the zappy gun does for electricity puzzles. By just providing power everywhere, the puzzle possibilities are simply gone.
So we gotta make sure abilities always only provide one part of a recipe, not the entire recipe at once.
Secrets
Secrets are very fun to find if they are not too easy and if they are all different in a way. Six Inches Under has a lot of tedious “look into every corner” and you can be sure there will be something everywhere. But none of them feel special to find. It just feels like work going through all of them, not like wonder. Especially in the industrial area I find it especially unfun to just go everywhere. Did you feel that too or is it just me?
In Supraworld all the secrets will be unique and special again, and I’ll try to keep the areas uncluttered. The clutter in the cave creates so many rocky corners you need to go through. We’ll have straight walls again from now on!
I think pretty much everyone agrees that the secrets in SIU are not that rewarding. We keep hearing complaints about “just coins/health”. It’s one of the hardest things to keep rewarding players in a meaningful way. Now I have a pretty good idea how to keep them motivating throughout.
A main problem in all of the three past Supraland adventures is that coins are absolutely necessary in the beginning to buy major upgrades; it’s super motivating and the balance is just right, but after that, coins become meaningless. You can buy some power ups, but you could also ignore them. I want to keep the upgrade loop with coins going throughout the entire game. So even a secret with coins should feel meaningful and exciting.
Supraland had about 210 chests, SIU had 140. It should be clear that we cannot give you something super meaningful in every chest. But with the right balance even the less exciting things can feel great to have.
But in Supraworld we will also give you really meaningful stuff in secrets, like new optional abilities. It should be really exciting if every chest is potentially one of these. I’m also thinking about randomizing chests a bit, so you never know for sure where things will be.
Endgame
I like how SIU has so much endgame content, like it has 3 more areas. In Supraland you’ve basically been everywhere when the credits rolled. You could find more stuff, but little entirely new. Having so much endgame stuff also meant, the credits roll much earlier and it’s a frequent complaint that I saw coming exactly that way. Even though there is a lot to be done, when you’ve seen the credits, you kinda feel, the game is over and the motivation drops.
In Supraworld I wanna have even more optional regions, but I want them accessible before the ending. So you’ll most likely be going places and be spending a lot more time before the credits.
Easy peasy
We keep hearing how the puzzles in SIU are too easy compared to SL.
I totally agree with that one. We had a tendency to help players through the game way too much, while I personally enjoy seeing people suffer while being stuck on puzzles. Not every puzzle should be hard, but every now and again. The joy is so much bigger when you finally find the solution after you were stuck for a while. So yes, I want to make Supraworld significantly harder again in terms of puzzles.
The puzzles in Supraland 1 also happened in more complex situations. There were more options that made the situation less clear. SIU often only leaves you one thing to do so it's obvious what to do.
I think developers are generally afraid of players facing too complicated situations and leaving the game for good. But I think the opposite is true. It engages players even more and they wanna figure out how to do it. The puzzles just need to turn out to be fair, or otherwise you lose the trust into the game.
Does that make sense?
Puzzles
When I built everything on my own, I was guided by my feeling for how puzzles should be done and how secrets should be laid out. When I brought more people into the process this naturally changed and everyone kinda felt “this doesn’t feel like Supraland”. So instead of just feeling the design of the game for myself I had to think harder about it and put it into words.
There were a lot of puzzles that we cut from the game or reworked entirely, and it got me thinking, why do these puzzles not work? I could feel the reasons, but I needed to put them into words for others to understand.
As a guiding system I came up with 13 points to check each puzzle against. It turned out to be super useful even for myself.
Not all the criteria must be met, but if you go through the list and compare any puzzle with it, it will certainly lay bare all the weak aspects of it. I think for the most part these points should be suitable for any puzzle in any game, not just Supraland.
13 Puzzle Rules - Checklist
1. Do the players set the goal for the puzzle themselves?
Best case is you set a goal for yourself and you were not told about it. The more agency, the better.
2. Does the puzzle add a new aspect or unexpected twist to a mechanic?
If it doesn’t, do you need the puzzle at all or is it just boring repetition? Also make sure the game doesn’t tell you about the new mechanic or it’s worthless to have it.
3. Players generally figure out the trick (if available) without being told about it.
Only subtle hints are allowed. Straight out being told how to do it by game, makes the puzzle useless.
4. Is the puzzle focused on one idea or are there too many at once?
If it can be cut down to its idea’s essence, do so! If there are lots of simultaneous little steps involved, it might be just convoluted and not fun.
5. Once you had the “AHA!” moment and you know the solution, is it reliably easy and quick to execute?
It’s not fun if you have to wait a lot or have to jank objects around until it finally works after many attempts.
6. Are all consequences of your actions predictable?
If done right, you don’t need to try lots of things, but you can solve the puzzle just by looking at it, because nothing unexpected will happen. Everything makes sense.
7. Player does not solve it by chance or by brute forcing it?
It has to be absolutely sure that the players understand exactly what they are doing and don’t stumble over a solution and they cannot just try a number of possible options until they got it.
8. Is it impossible or very unlikely to find the solution before you even encounter the problem?
For example: there is no fun to be had if you find a key before you got stuck at the locked door.
9. There are no wrong solutions that players keep on trying?
Eliminate them! If people can almost solve the puzzle in the wrong way, they will keep trying that forever. If a jump is not supposed to work at all, make sure it’s clearly impossible. If it’s almost possible, you’ve got a problem.
10. Does the puzzle NOT open a door and NOT help you get to an elevated place?
If you can give rewarding progress in another way, that’s great. Progressing via verticality or unlocking a door should not be the only thing you always do.
11. Avoid portal-test-chamber things like buttons, jumppads and stuff.
Buttons that hold a door open or move a platform or so make no sense in a real world and only Portal is really ‘allowed’ to use them. These are mechanics that have “Portal-test-chamber” written all over them and feel wrong to me because they don’t feel like they should actually be part of a game world. We don’t need any more games where you need to put cubes onto buttons.
12. Can the player be surprised by seemingly unrelated things being part of the puzzle?
If you suddenly notice that some element of the world is part of the puzzle, that’s a great revelation. For example if an omnipresent decorative thing is being used for the solution.
13. Can its red herrings, decorations or puzzle elements be used again for another puzzle?
It’s always great if all the parts of a puzzle have more than just one use.
Examples
Let me go through this list by looking at my favorite puzzle I made: the halo puzzle in Supraland 1.
You need to help a guy get into a chapel. He is not allowed entrance because he is not holy enough. The process is that you realize he probably needs a yellow halo over his head to be allowed in. You find a fitting shape but it has the wrong color. You then figure out how to color things by yourself.
This list will show the strong and the weak parts of it:
1. Yes. The NPC tells you what he wants to do, but he does not tell you what you need to do to help him. By figuring out the problem on your own, you then automatically set further goals for yourself.
2. Yes. There are new mechanics. The puzzle introduces the entire color puzzle mechanic into the game.
3. Yes. Everyone figures out all the important parts by themselves. Not too hard, not too easy.
4. Yes. The puzzle is very straight forward. It’s basically 2 clearly separated steps; nothing convoluted.
5. Yes. Once you know what to do, the execution is very simple and even fun because it’s a new and unique thing to do.
6. Yes. The consequences of everything are fairly predictable. You will definitely experiment a little but everything makes sense within the game's logic.
7. Yes. You will absolutely not randomly solve this.
8. Yes. Impossible to stumble over the solution before seeing the problem.
9. Yes; all the wrong options are clearly wrong. There is one typical wrong thing that most people try at first (a rusty, a white or a red halo) but there is no reason to try it again.
10. No, it ultimately opens a door for progress. This is a weak spot I feel since I made it. He simply opens a door for you once he enters the chapel, and that is pretty lame. Once in the chapel he could have given you progress by for example using the power of the church to change beliefs of other NPCs, so they would have given you new options to act. The simple physical/geometrical progress lacks creativity.
11. Yes. There are no articifial elements to the puzzle that seem out of place in the world.
12. Yes. Unseeming decorative things become puzzle elements multiple times here. At first you learn that flowers can be put into the color machine to absorb their white color. Then you need something yellow though and there are no yellow flowers anywhere. The only thing yellow are the crystal lamps everywhere. People are always like “but there is nothing yellow” and then they get super excited to find out they can take the crystals out. They didn’t notice them even though they’ve been there all along in many areas of the game. There is even more decoration in this place that later becomes meaningful.
13. Yes and no. Some elements of the puzzle will be used for other things as well, some don't. I added red herrings: the shabby store in the area gives away rusty objects of all shapes. This is to disguise the rusty ring by also having a rusty cube and cylinder. These other rusty objects have no further use, which is a missed chance. But the color machine will be used for up to 3 more different things right after which I think is fantastic.
So this puzzle fulfills 11.5 out of the 13 points.
Good puzzles usually check around 8-10 of the boxes, Anything above that rarely happens.
If a puzzle only checks 5-7 boxes, you should really consider removing or improving it.
More Examples
If you take a very early puzzle from Supraland 1 for example: getting the golden barrel that has the double jump upgrade inside: you need to roll the barrel through a small gap out of a hut and prevent it from rolling into the lava by placing the force cube in its way. It scores a 10/13. It is a very basic little puzzle, but you can feel people’s excitement when they work it out.
In Six Inches Under we had some 5-7 point puzzles that we improved a lot or replaced entirely. You could just feel they had nothing exciting about them. A frequent problem was that the consequences of your actions were totally unclear, so you unintentionally solved problems and just stumbled forward that way. There was also just no “Aha!” moment, which is the pivotal point of any puzzle. Without that, it doesn’t need to exist.
A puzzle in SIU, that I really like, fulfills 10 points: dropping the battery from the tower to get the racers to start running. It’s only held back by how it uses some typical portal-test-chamber like objects and the goal is to gain more verticality.
The last serious puzzle in the main quest of SIU is in the boss castle, where you need to float the match to the battery. Everyone agrees, it’s a great puzzle! But it checks 'only' 9 boxes. What's holding it back from real greatness are the 2 activators for the translocator and the force cube and that you can try some janky, weird stuff that can eventually also solve it for you in a completely wrong way.
There were ways you could have solved it with the translocator or the Forcecube (by teleporting to places or by jamming doors) so we just had to force you to park these 2 abilities somewhere to make sure you solve it in the most beautiful way. These additional elements just add so much noise to the puzzle so that you cannot clearly see the beauty of its essence anymore.
Ideally the puzzle should have been in a place where you didn’t have these 2 abilities yet; but that wasn’t possible because of the special puzzle mechanic in the boss castle, that was also required.
A SIU puzzle that fulfills 7 out of 13 points is getting through the dark cave with a hot glowing metal ball. It’s kinda cool but it lacks in the agency department. It does not fulfill the first, third and seventh point on the list. An NPC kinda tells you what to do and you don’t know what you are doing it for. So the rest sort of just happens, but you didn’t plan for that. You kinda solve it by accident. On the plus side it got the points for not using typical test-chamber stuff, its elements can later be reused for other puzzles and you are not progressing through a door. Instead the progress happens by lighting up the path forward that you simply could not see.
What I found really interesting, there are 3 puzzles in SIU puzzle that tick 12 out of 13 boxes:
1. The optional one at the beach, involving the beach guard, the showerhead and the lady. We had to create 3 entirely new mechanics just for that one...
2. "Making the blue key card", which I think is beautiful, also the way we hint at it.
3. The whole thing about getting the lady to smell like flowers.
Which puzzles gave you the best experience? Please let me know, I'm super curious.
How to come up with puzzles anyway
If you are here, because you’re interested in the puzzle design process, rather than SIU, I'll outline my process. The previous bit was about checking if a puzzle works or which issues it might have. But I always notice, when speaking about the process to non-puzzle-designers, there is big confusion about how to even get puzzle ideas.
Players think through a puzzle the exact opposite way that the maker does; I think that’s why it seems so hard. Making it can be much easier, most of the time.
My process of finding puzzle ideas is all about thinking of “what things can you do with those elements and abilities”. Once you found a mechanism, you just try to build something that forces the player to use that mechanism that you just found and avoid any other possibilities. On the way you will run into lots of problems from the 13 point list above.
The origin of the puzzles is mostly just playing around with the things you created and you suddenly do something funny that gives you a new idea. That is what makes it so hard to plan it all out ahead. Most good ideas only come to you during development. Then it’s sometimes hard to keep everything consistent because you keep adding new interactions between objects, and suddenly you break other puzzles because of the new options you created.
All of the color mechanics of Supraland 1 were improvised into the game. The whole blaster mechanics of shooting through the beam/colors/electricity etc were shoehorned in very late as well. While shooting my beam with my blaster I just felt "it should become purple now". I planned out the Supraworld abilities (all new) in detail, but I'm curious what kind of new ideas we'll get along the way.
For the sake of it, this is the puzzle that was among the hardest for me to create.
The start of the process was, while playing around during testing, I noticed I can remotely move those steel beams back and forth with my force beam without reconnecting anything in between. I found the idea cool to remote control something, and that the players need to set it all up by themselves.
The rest of the process is finding a way to make that remote control work. It could have been something to jump on, something that blocks your path or whatever you can come up with. I went for creating different electrical connections on each side of the beam to control different doors.
But why is it all in a glass cage? The Zappy gun just got in the way. You could have shot the antennas, it would have made it so easy, you would not have noticed there was a puzzle. So with the variety of abilities you need to exclude all of them from the equation.
There kept being more and more alternative ways to solve it, that were all way too boring so I kept moving stuff around and blocked off more and more until it worked. The magnet under the glass box is also part of preventing a fake solution, that involved bringing an anvil into the building.
Now that I think about it again, I guess I could have made the cables fill slower, so the zappy gun alone would not have been enough because it doesn't last. But when I made it we did not have slowly filling cables. That mechanism was only added months later.
Supraworld
I could go on with things we learned forever and we all can't wait to finally get started with Supraworld and leave SIU behind. The whole team will move over on the 31. of January 22.
Some of us have been pretty busy with Supraworld already, mostly creating assets, and I started building the game world. But there is no coding done yet, so there is nothing functional. If you haven't seen it, here is a very early glimpse.
https://www.youtube.com/watch?v=IF6YPgH_XY0
I would like to go Early Access with Supraworld pretty early actually. It was great in the first Supraland to have so many testers play the game in Early Access from the very early versions on. It's very helpful in terms of feedback and motivation. If you join it, just don't expect anything close to a finished game. You'd buy in to being a spectator to the entire development process.
I would love to read all your thoughts on anything here. I'll read all of it!
For the end I have this little strawpoll about a question I'm really curious about: Poll: Which Supraland game did you play first?
- David Münnich